New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Seems the NVMe box in Chicago still have issues, network and IO are poor and unstable, and the internal network disappeared again, reconfigure network in the panel not working this time, always show an error "Something went wrong. Please try again in a few minutes."
me too.
So do I
Me too. Now I temporarily use the public network connection. The speed is very, very slow
my two CHI storage boxes are fine
We're doing our best to get things up and running asap for the affected customers.
I am curious to know how you use this as backup since it has very limited CPU. My Amsterdam one has been idle since I got it. Ideally, I want to use nextcloud but I am afraid it will use more resources than permitted.
I created an unprivileged user account called
rclone
on this HDD server.Elsewhere, I can use rclone SFTP remote targeting this user account.
Content in my Seafile file share is uploaded daily to this server, using a cron job on the machine hosting Seafile.
This backup is encrypted via rclone crypt remote.
I did a restore test shortly after initial setup.
My family photos are uploaded daily, using a cron job on the OpenWRT router.
Given the router has only 128MB RAM, I need to restrict memory demand with
--buffer-size 1M --max-backlog 64
flags.This backup is not encrypted.
I can verify by looking at the target folder.
May I ask why you choose SFTP rather than WebDav with https? I think WebDav may have better Concurrency.
Benchmark junkies are the bane of humanity.
The backup runs every 24 hours.
As long as it completes within 24 hours, before the next backup begins, it is a successful setup.
SFTP is easy to setup and authenticates with a private key.
WebDAV is more complicated (it requires TLS certificate) and authenticates with password (less secure).
Simplicity wins over performance.
Thanks I understand.
Dear Hosthatch team, my 2T hard drive server in Chicago has been disconnected for more than 24 hours, and still can't connect, I submitted a ticket in the backend a day ago, and no staff replied, but my friend have the same IP segment of the same configuration machine can connect normally, please ask your side of the network has not been repaired?
The ticket number I sent is #####263066 If you can, please help solve it as soon as possible, I just renewed a year yesterday, and there are some services on the machine need to be restored as soon as possible, thank you!
I use Borgbackup on mine. Most of the work for Borgbackup is done on the client-side, so it's not too CPU-intensive on the server.
I'm using Seafile, but I'm running Seafile on an NVMe VPS and mount the storage VPS via NFS. It works great! Seafile is very fast since its database is on a fast NVMe disk, while the actual large files are on the storage VPS.
I wouldn't recommend SSHFS or Samba for Seafile as it relies on hard links (which Samba doesn't support, and SSHFS sometimes corrupts), but NFS and iSCSI both work well.
the internal network in Chicago backed to normal now, but the Nvme IO still badly
I did the same as you, Nvme's machine is now working fine after maintenance, but unfortunately my 2T machine is now unreachable, resulting in many of my services not working properly @hosthatch
Hopefully lessons learned for the other locations.
Anyone else noticed that the Ubuntu 18.04 template is broken on the storage? I reinstalled on one of my 2T boxes, but it only came with 2G of space.
You just need to expand the partition.
I was so happy with my 8GB Chicago NVMe on Nov 11th
https://i.postimg.cc/4dh6p7tJ/Host-Hatch-Chicago-2021-11-11.png
Today:
https://i.postimg.cc/pT8jmNbr/Host-Hatch-Chicago-2021-11-17.png
Now I need to decide tonight if I should renew it. (I have account credit)
@hosthatch if I cancel and it improves can I remove the cancellation before the expiration date of Dec 2nd?
EDIT: I am going to have faith that it will get resolved.
My chicken has hatched!
Have you looked at MacSec for the private link?
Only getting like 300Mbps using NFS over wireguard and Resilio.
I've never heard of MacSec before! Something new to explore. I'm not sure how well KVM would handle a custom layer 2 protocol (if it even supports it), but there's no harm in trying it out. Might be worth trying IPSec too.
Were you testing the transfer speed on these newer EPYC processors, or the older Xeons that HostHatch normally use? With any sort of encryption, you very quickly hit speed limitations in the legacy Xeons, even with a hardware-accelerated cipher like GCM-AES-128. The newer Xeons and EPYCs are noticeably faster.
You can run this:
And it'll give you the raw speed of that encryption algorithm, which is the max you're going to reach over any encrypted connection that uses it.
Having said all that... WireGuard uses ChaCha20-Poly1305 which isn't even hardware-accelerated, and I get 1.6 Gbps via WireGuard over HostHatch's internal network in Los Angeles:
I get around 5 Gbps when unencrypted.
If you're using HostHatch's internal network, make sure you enable jumbo frames (on the internal network only) by setting the MTU to 9000, eg:
I think I was only getting around 200 Mbps without jumbo frames.
WireGuard will automatically adjust its MTU based on the underlying network adapter's MTU.
"Promotional Package HDD 1TB"
cat /proc/cpuinfo | grep name
model name : Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
Oh, I forgot to mention (and it's too late to edit my earlier comment) that I usually hit IO limits before network or CPU limits on the HostHatch storage VPSes. Disk IO can be slow at times, with high iowait - Noisy neighbours maybe? Last time I ticketed that, they said that they didn't see any iowait on their end. Having said that, I really like HostHatch because the storage VPS is really good value for money, and it's definitely better than what I expected for the price. tbh I'm not sure how they can offer what they do for the price they do.
Patiently waiting for ams upgrade. Not sure sg would be upgraded since it was launched recently.
The recent EPYCness of HostHatch is preventing me from cancelling my idle AMS NVMe server. It would be impossible to get a 40G EPYC deal in AMS from such a reputed provider. Singapore is still good with the current CPUs for the time being. Also, they offer more BW than many other providers there.
@hosthatch Would the dedicated CPU usage allotment values (such as 12.5%, 25%, etc.) on various previously sold plans/deals change with the ongoing upgrade to EPYCs?
Yep, totally agree. I'm very happy with my storage and NVMe servers from HH.
When is Stockholm's turn?
haven't received any notifications yet, and guess need to wait for a couple of months.